Table 1 Computational methods for prediction of TAD hierarchy

Approach	Caller	Input format	Main language	Parameter
Liner Score	Arrowhead	hic format	Shell, Awk, Java	1
	Armatus	dense matrix, sparse matrix, Rao format^*	C++, Python	1
	CaTCH	catch format^**	C, R, Shell	0
	HiTAD	cool format^***	Python	1
	matryoshka	dense matrix, sparse matrix, Rao format	C++, Shell	1
	OnTAD	dense matrix, hic format	C++	2
	Multi-CD	dense matrix	Matlab	NA
Clustering	IC-Finder	dense matrix, sparse matrix	Matlab	NA
	TADpole	dense matrix	R	6
	BHi-Cect	Rao format	R	0
	SpectralTAD	dense matrix, sparse matrix, hic format, cool format, Rao format	R	3
Network features	HBM	dense matrix	R	5
	spectral	mat format^****	Matlab	NA
	3DNetMod	sparse matrix	Python	18
	GRiNCH	sparse matrix	C, Python	3
Structural Entropy	deDoc	sparse matrix	Java	0
Structural Entropy	SuperTAD	dense matrix, sparse matrix	C++	0
Statistical Model	TADtree	dense matrix	Python	6
	GMAP	dense matrix, sparse matrix	R	4
	PSYCHIC	dense matrix	Matlab, C	NA
	HiCKey	dense matrix, sparse matrix, Rao format	C++	6

A matrix is a two-dimensional data object made of m rows and n columns, therefore having total m x n values. If most of the elements of the matrix have 0 value, then it is called a sparse matrix. In Hi-C data, the sparse matrix represents the chromatin contact map, the numerical values in row i and column j represent the frequency of DNA interaction between i bin and j bin in chromosomes. The sparse matrix is one of the common inputs for TAD hierarchical structure recognition algorithms.
^*Rao format is another sparse matrix, of which the start and end sites are represented by genomic coordinate.
^**catch format, ^***cool format, and ^****mat format mean files separately produced by CaTCH, cooler, and Matlab.
Bold text is the actual input type for this article.

Quick links

Search